 |
I recently received a copy of the Mathematica computer mathematics software for the Apple Macintosh II from Wolfram Research. To see if it has potential for application to artificial intelligence at the U. S. Geological Survey, I applied Mathematica to a problem in neural network computing. I selected the interactive activation and competition (IAC) neural network model as developed by James McClelland and David Rumelhart of the Parallel Distributed Processing (PDP) Development group. The theoretical basis for the IAC model is discussed in detail, along with other neural network models, in their Parallel Distributed Processing textbooks. Practical exercises that relate to understanding the PDP neural network models are described in their handbook, Explorations in Parallel Distributed Processing. I used the description of core routines in the PDP handbook to develop a Mathematica version of the IAC model. This article outlines the results of the evaluation.
|
|
For the newest resources, visit Wolfram Repositories and Archives »
